← harvey / Staff Product Manager, Agent Platform
tailored_resume_v2 / art_ggnDPVg7CVw
role
model
anthropic/claude-sonnet-4.6
created
2026-05-21T22:44
↓ Download .docx ↓ Download .pdf PDF requires LibreOffice installed
What changed for harvey
| change | why it matters |
|---|---|
| Summary rewritten to lead with '0-to-1 AI products and enterprise-grade platforms' and explicitly name OpenClaw multi-agent orchestration | Harvey's first requirement is 0-to-1 product experience and the role is explicitly an agent platform — OpenClaw is the strongest direct proof point |
| Summary embeds 'trust systems, governance infrastructure, agentic AI, enterprise complexity' from JD | These are the core differentiating requirements Harvey calls out beyond standard PM experience |
| Streamio reordered to lead with OpenClaw multi-agent orchestration bullet instead of Electron app bullet | OpenClaw directly mirrors Harvey's agent platform architecture; it is the single strongest proof point for this role |
| Streamio bullets reframed to emphasize agent interaction patterns, natural language workflows, domain-specific agent scoping, and 0-to-1 product execution | JD's core surface is 'how lawyers interact with agents, how agents get created and discovered' — OpenClaw and StreamIO's agent workflows map directly |
| Fintellect bullets reframed to emphasize structured output validation as 'trust and verification infrastructure' and domain-scoped agents with firm-level constraints | JD explicitly calls out trust systems, verification flows, governance infrastructure as core product requirements |
| Intuit drift detection bullet reframed as 'governance and compliance infrastructure at enterprise scale' | Harvey's JD states ethical walls, audit trails, governance 'aren't afterthoughts — they're the product'; Intuit's drift detection maps to this |
| Splunk bullets reframed to foreground audit trail, governance, and compliance requirements of Splunk Cloud | Harvey's enterprise complexity requirements (ethical walls, audit trails, multi-stakeholder flows) map to Splunk Cloud's regulated enterprise context |
| Kaiser Permanente first bullet reframed to explicitly call out 'regulated healthcare environment' and 'governance, audit trails, multi-stakeholder approval flows were core product requirements, not afterthoughts' | Mirrors Harvey JD's exact language about governance not being an afterthought |
| Projects section reordered to lead with aeval instead of RL Workbench | aeval's safety testing, refusal detection, automated safety gates, and audit infrastructure map most directly to Harvey's trust/governance/ethical walls requirements — stronger signal than RL benchmarking for this role |
| aeval bullets reframed to explicitly connect adversarial safety testing and automated safety gates to 'ethical walls and audit trails' | Harvey JD calls these out as the product, not features — aeval demonstrates Felix has already built this infrastructure layer |
| RL Workbench second bullet reframed to connect technical depth to 'engaging with engineering on frontier agentic AI architecture tradeoffs' | JD requires comfort with technical depth and ability to engage engineers on system design — 12-algo RL implementation is the strongest proof |
| Bank of America role removed from experience section | 1-summer internship from 2011 adds no signal for this role; space optimization for 2-page target |
| AutoEval condensed to 1 bullet and moved to last project position | Least directly relevant project for Harvey's agent platform role; space optimization |
| Teaching experience condensed to 1 bullet listing all courses | Not a differentiating signal for this role; space optimization for 2-page target |
JD analysis (19 key phrases)
Key phrases: 0-to-1 product workagentic AIagent platformmulti-agent orchestrationtrust systems, verification flows, and governance infrastructureethical wallsaudit trailsenterprise complexityend-to-end product lifecyclelawyers interact with agentsagent creation and discoverydesign partnershands-on with the detailshard prioritization calls with imperfect datafirm-level deployment constraintsmulti-stakeholder approval flowsfrontier agentic AIenterprise-grade platformnatural language workflows
Hard requirements:
- 7+ years product management experience
- End-to-end ownership of user-facing product (not just features)
- 0-to-1 product building or major product pivots
- Strong UX product instincts
- Technical depth — system design, architecture tradeoffs
- Enterprise shipping with governance/compliance/regulatory constraints
- Operate in ambiguity, form strong opinions quickly
Preferred qualifications:
- Experience with AI/ML products
- Legal technology experience
- Professional services domain knowledge
Per-role mapping (11 roles scored)
| role | score | reframe angle | JD phrases that map |
|---|---|---|---|
| Streamio AI — Founder & CEO | 4/5 | Multi-agent orchestration platform builder — lead with OpenClaw and agent workflow architecture, frame real estate/insurance/financial agents as domain-specific agent deployment analogous to legal | 0-to-1 product work, multi-agent orchestration, agentic AI, agent creation and discovery, hands-on with the details, end-to-end product lifecycle |
| Fintellect AI — Founder & CEO | 3/5 | Domain-specific AI agent platform with structured output governance — frame as enterprise-grade AI product with compliance-adjacent validation | agentic AI, trust systems, governance infrastructure, design partners, 0-to-1 product work |
| Intuit — Staff Product Manager | 5/5 | Enterprise platform PM who shipped 0-to-1 infrastructure products at scale with measurable adoption and revenue impact — frame governance/lifecycle management as analogous to Harvey's trust/audit infrastructure | end-to-end product lifecycle, enterprise complexity, multi-stakeholder approval flows, hard prioritization calls, hands-on with the details, firm-level deployment constraints |
| Splunk — Senior Product Manager | 4/5 | Enterprise platform PM with audit/governance experience and fast 0-to-1 delivery — Splunk Cloud's compliance requirements map directly to Harvey's ethical walls and audit trail needs | audit trails, enterprise complexity, governance, hard prioritization calls with imperfect data, end-to-end product lifecycle |
| Kaiser Permanente — SOA Technical PM | 3/5 | Regulated enterprise platform with compliance and multi-stakeholder governance — condense to 2 bullets emphasizing regulated environment and enterprise scale | enterprise complexity, governance, multi-stakeholder approval flows |
| IBM — Software Engineer | 2/5 | Technical foundation — condense to 1 bullet | — |
| Bank of America Merrill Lynch — Tech MBA Associate | 1/5 | Condense to 1 bullet or cut if space-constrained | — |
| RL Workbench | 3/5 | Demonstrates AI/ML technical depth required to engage engineers on architecture tradeoffs for frontier agentic AI | frontier agentic AI, technical depth |
| aeval — AI Model Evaluation Platform | 4/5 | AI governance and safety infrastructure builder — lead project with aeval to signal trust/verification/governance instincts | trust systems, verification flows, and governance infrastructure, audit trails, ethical walls |
| AutoEval | 2/5 | Condense — shows AI evaluation depth but less directly relevant | — |
| BRAIN — Protein Structure Prediction | 2/5 | Condense — NeurIPS credential supports AI/ML bonus qualification | frontier agentic AI |
Tailored summary
Technical Product Leader with 12+ years shipping 0-to-1 AI products and enterprise-grade platforms at scale — from building multi-agent orchestration frameworks and domain-specific AI agents (OpenClaw) to scaling platform infrastructure to 675M+ engagements at Intuit. Deep instincts for trust systems, governance infrastructure, and the hard enterprise complexity that makes agentic AI actually deployable. NeurIPS published AI researcher; hands-on across the full stack from RLHF post-training to production agent workflows.